| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We've seen the NFCC get in a state where
it continuously reports field off/field on
events. The danger is that the device goes
to sleep right after a field on event. On
wake-up, the field off/field on loop is
gone, and the device prevents any access
to the SE from the DH because it thinks
the field is still present.
Bug: 7386840
Change-Id: I87b681124078fe69c0244efd7b14ebe3ffd1c161
|
|\
| |
| |
| |
| | |
* commit '547b58f41866671e0028ed6f395b34e271ce81b1':
Fix tag not deactivating to sleep during reconnect.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
During reconnect, the JNI deactivates the tag to sleep mode
and re-selects it. Sometimes when the tag leaves the RF field, it
deactivates to idle instead of sleep; causing subsequent select to fail.
Author: Mark Gorodetzky
Bug: 7357299
Change-Id: Id774d0bc3b8e2cbcea93cd2ee061ffecfa821a1e
|
|\ \
| |/
| |
| |
| | |
* commit 'a15b24e9e405b65b7523d8c80173b1c0d71e093e':
Check for invalid handle before configuring secure element.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If handle is invalid, don't adjust secure element's listening
parameter. This prevents a warning message from stack.
Use LLCP_MAX_MIU to prevent a warning message from stack.
Bug: 7357666
Change-Id: I79af30f7d3db1d597b46d8cf43991c9157359e10
|
|/
|
|
|
|
|
|
|
|
|
|
|
| |
We used to acquire a timed wakelock when either resuming
or suspending. We don't want to take a timed wakelock
when suspending, since we want to suspend asap, but also
don't want to underlock the wakelock.
Instead of taking a timed wakelock, just abort
the process in case the NFC controller hangs.
Bug: 7341879
Change-Id: Ibd2f468264f60a39db3cb271a6f05c50dfb559bf
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Users of the NFC-extras API usually want to talk
to the SE once it has completed a transaction.
The hard part is knowing when it is safe to connect
to the SE, as it will break the connection to any
reader. And when relinquishing the connection to
the SE, the reader may find the device again and
process another transaction.
This patch adds the following two conditions
for allowing to open the SE from the DH:
1) The SE may not be activated in listen mode
2) The RF field must have been off for at least 50 ms
Bug: 7275484
Change-Id: Ibde32a8e2aef045c17ab76ef08c72f96edfedaef
|
|
|
|
|
|
|
|
| |
Create the RF Events for SE transactions based on RF
activations (listen mode and ISO-DEP only).
Bug: 7276275
Change-Id: Ibea0f64ba82ed31a7c582f6a44eb15abf25dda1c
|
|
|
|
|
|
|
|
|
| |
Causes applications such as TagInfo to report the tag
as being lost. Also fixed a small memory leak triggered
in this case.
Bug: 7294613
Change-Id: I260e02896e534b23de5c7e0278609e27a9a20e7f
|
|
|
|
|
| |
Bug: 7282057
Change-Id: Ic877f67654906e21f1942adf05751943e3d6cb01
|
|
|
|
|
|
|
| |
Protect state against multi-threaded access.
Bug: 7281263
Change-Id: I12735c725b6abf956cb7143f25397653839bc6e4
|
|
|
|
|
|
|
|
| |
Workaround some SEs that do not support interpreting
an empty select as a select of the card manager.
Bug: 7198430
Change-Id: Icce0e793ca6f03d5c2271812859b04417d2708e8
|
|\ |
|
| |
| |
| |
| |
| | |
Bug: 7264008
Change-Id: I2647b0c6adf8f95e21d9f789c411742aef19c5ad
|
| |
| |
| |
| |
| |
| |
| |
| | |
We need to allow usecases were the DH wants to talk to
the SE even if the screen is off.
Bug: 7257431
Change-Id: Iad66ef29757985b7aab5f0748fbf7589b3777bc3
|
|/
|
|
|
|
|
| |
Could cause RF field to remain on.
Bug: 7164846
Change-Id: I82efc29b96c7129465fea5354fa5484da31618a6
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If we leave RF polling/listen enabled while talking
to the SE from the DH, there is corruption on frames
sent out over SWP. To resolve this, disable discovery
(both poll and listen) when the SE is route to the DH.
This also makes the behavior identical to that of
the PN65N devices, which don't allow any RF activity
when the SE is routed to the host.
Bug: 7004303
Change-Id: I05ad55a6f75b9f346feaf20dccefc23919ff538f
|
|
|
|
|
|
|
| |
Fixes coming out of suspend with EE routing enabled.
Bug: 7233312
Change-Id: Ibbb9d7c26c2152fbac543d84d3306fc0ed50272d
|
|
|
|
|
|
|
|
|
|
|
| |
Set HCI registry to cause Oberthur secure element to perform warm-reset.
We use this to close any applets that the DH may have opened during
communication to the SE.
This feature is Oberthur-specific.
Bug: 7163680
Change-Id: Id0589b5b41bbea1df9b8dbd73abd5a112f773916
|
|
|
|
|
|
|
|
| |
When NDEF detection times out, stop polling temporarily.
Let NXP PN544 controller poll and activate the NFC-DEP interface.
Bug: 6986507
Change-Id: Iaa3194e4dbc4b9d2c16aaeb360a7f15dfd82c38a
|
|
|
|
|
|
|
| |
Would otherwise cause stack hang.
Bug: 7214014
Change-Id: I5a5373f389b06b873cee1c1fb694502d7ad598b3
|
|
|
|
|
| |
Bug: 7113095
Change-Id: Iaa2e2b05244ec10392c98aa5dd0d5ee84e62e39f
|
|
|
|
|
| |
Bug: 7125646
Change-Id: If700eb3cdd85e9ee08701d2dd8971ce2500a834d
|
|
|
|
|
| |
Bug: 7125646
Change-Id: I4de457f982857be315d65386f431d5eafcd8af02
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The PeerToPeer class had some serious multi-threading issues:
- mServers / mClients simultaneously being accessed by different threads
- P2pServer / P2pClient objects being deleted while still being used by
other threads
- Race conditions where we could go to sleep on a condition variable
without ever waking up.
This fixes most of the issues by:
- Properly locking all access to mServers / mClients
- Using Android's sp mechanism to do ref counting on P2pServer/P2pClient/NfaConn
TODO:
- Fix races around the condition variables - those are very hard to hit,
but we need them fixed anyway.
Bug: 7089048
Change-Id: I08564c65cc40734086e517d6107e31f86c3b6adc
|
|
|
|
|
| |
Bug: 7123942
Change-Id: I4fa153b80e1af2b00b326370e102f714cde4334b
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
When reading a tag and it does not respond to a command, use taglost=1.
It causes a TagLostException in the app.
Bug: 7047415
Change-Id: I766849991441b82e5ad03d3031bc735fa5f9c27a
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| | |
The "app name" parameters must match in HciDregister() and
HciRegister()
Change-Id: I43010268bb4c0ee7cdc7c1440346a6dedc230a58
|
|/
|
|
|
|
|
|
|
|
|
| |
At the end of nfcManager_doInitialize(), enter low-power mode.
This logic is used after boot and before the user unlocks the screen.
If the user does not unlock the screen, the phone stays in low-power mode.
Additionally, only power off the chip if discovery is disabled
AND the SE is not selected.
Change-Id: I2bd2effce34b5dc586ad484c1da2ae86da49f0c1
|
|
|
|
|
|
|
| |
Use the poll duration ("NFA_DM_DISC_DURATION_POLL")
from the conf file if one is configured.
Change-Id: If17d402cbe8505b096a168640f99b88f7e3b6e48
|
|
|
|
|
|
|
|
|
| |
This code was used when Android did not support
SNEP natively yet. The current Android behavior
is to automatically try SNEP first, then fall back
to NPP.
Change-Id: Ifb1df3a5439cc1a7b939b4ae1dc70135d9c63a25
|
|
|
|
| |
Change-Id: Ie747c06a0a5161ad5479f93cbc56e482ed7da33c
|
|
|
|
|
|
| |
Correctly handle NFA_P2P_CONGEST_EVT to prevent race condition.
Change-Id: If255ade96867d45b2a7cff1cb3d6e4cbc2d883d3
|
|
|
|
|
|
|
|
| |
BLTH01652900: During trasceive operation for Mifare Ultralight, handle
NACK and ACK responses properly
Bug: 7047415
Change-Id: I778a2bbbcd3fff47b0da73ec37d76bf81708d931
|
|
|
|
|
|
| |
BLTH01642677. Remove "brcm" in names.
Change-Id: Ic3a08abfeb2808d2a0b01402a76d9864b099ef49
|
|
|
|
| |
Change-Id: I79510a45e8a552626f0473000e23d0fc6888ee33
|
|
|
|
|
|
| |
Create JavaClassConstants.h to store some global names
Change-Id: I918d9e8f6db646f637cddbebbc2c4e5c5495623d
|
|
|
|
|
|
| |
In SyncEvent class, only unlock mutex once.
Change-Id: I0fb797c9e6e8ea793fd38abd72a0c2780a734e95
|
|
|
|
|
|
|
| |
BLTH01634648: During nfcManager_disableDiscovery() go to
power-off-sleep mode when secure element is not selected.
Change-Id: I85c37fa635a90fdc8c90b5a86fe1d3e091d421d4
|
|
|
|
|
|
| |
ALOGD is now squelched by default. Moved it to ALOGE.
Change-Id: I3bb7601e65d01d8d01a95902a0be95c7e1a3426f
|
|
|
|
|
|
| |
Remove unused functions; correct spelling
Change-Id: I6ecf17f8cf8eeae831ba3fcc0a2eb492332c0cb6
|
|
|
|
|
|
| |
Remove gJniVersion. Add thread-synchronization for obtaining a new JNI handle.
Change-Id: I23e05c95fd90005f6ef1cac69e1e07f0b9612227
|
|
|
|
| |
Change-Id: Iba0dab66be054ca1b0bc0f477c1107909df65cda
|
|
|
|
|
|
|
| |
- Logging changes
- Upgrade to stack NFA_MI_1.03.33
Change-Id: Iba2af1af49a261de50f79572b478838e2bdc890f
|
|
|
|
|
|
|
|
|
|
|
| |
Allows selectively enabling / disabling p2p modes in the
NCI stack.
TODO:
- Support selecting individual Nfc-F speeds (212/424).
- Support disabling initiator Nfc-A / Nfc-F just for NFC-DEP.
Change-Id: I55a08bd83640b45e2d8346587031f49213a0860b
|
|
|
|
|
|
|
|
|
| |
This wasn't directly referencing libastl, so I didn't find it first
time round. A top level rebuild without external/astl caught it. I've
since grepped and there are no other projects that pull in the include
files without the shared library.
Change-Id: I2dd06727098744076a7e115d05a8139029ff977a
|
|
|
|
| |
Change-Id: I0ff7c460b53f56d7b98e1e27838cfd4243a2b0a0
|
|
|
|
|
|
|
|
|
|
| |
In our current model these AIDs should be handled by
the EE.
Because this routing also activates Nfc-B listen mode,
it causes some Beam inter-op issues with PN544-based devices.
Change-Id: I865bbf7cdd48a781410ad3742426fa9b05751c70
|